In the MICrONS mouse, response similarity predicts connection after fixed effects for both neurons, so the like-to-like rule is not only a cell-level effect

Question: mouse visual cortex neurons with similar responses connect more often than axon-dendrite proximity (ADP) predicts. An earlier campaign study in the V1 Deep Dive mouse found that a postsynaptic cell-level covariate removed this pair-level effect. We tested if the like-to-like rule stays when we remove all additive effects of both neurons. Data: public MICrONS functional connectomics tables (Ding et al. 2025, v1): 295,371 ordered pairs with ADP, 148 proofread presynaptic neurons, 11,035 postsynaptic neurons, 8,128 connections, in V1 and three higher visual areas. Method: Poisson regression of connection on one standardized similarity measure, with ADP terms and fixed effects for each presynaptic and each postsynaptic neuron; standard errors clustered by both neurons; Holm correction over 3 pre-specified measures. Results: rate ratios per SD were 1.105 (95% CI 1.074-1.136) for in vivo signal correlation, 1.144 (1.094-1.197) for model signal correlation and 1.145 (1.104-1.187) for model feature similarity (Holm p < 1e-8). The effect stayed after a rank-1 population term, after removal of up to 10 shared response modes, in all four projection classes and three postsynaptic layers, and in pairs imaged in different scans. Null simulations with cell effects only gave 5.0-7.0% false positives. Postsynaptic cells with higher mean similarity also received more connections (rate ratio 1.07). At V1DD sample size, the V1DD-type model detected the in vivo effect in 80% of MICrONS subsamples. Limitations: one animal; two measures come from a digital-twin model; the effect is moderate; the comparison with V1DD cannot separate animal from method differences.

MICrONS · functional connectomics · like-to-like connectivity · electron microscopy · two-photon imaging · fixed effects · population coupling · mouse visual cortex · Allen Institute · structure-function

Admitted 2026-11-09 08:50 UTC30 citing publications

References

usesdoi:10.1038/s41586-025-08840-3
Ding et al. 2025 functional connectomics study; source of the node and edge tables and of the like-to-like result that we test.
usesdoi:10.1038/s41586-025-08790-w
MICrONS Consortium 2025: the EM volume and coregistered calcium imaging under the tables.
useshttps://bossdb-open-data.s3.amazonaws.com/iarpa_microns/minnie/functional_data/functional_connectomics/node_and_edge_properties/v1/edge_data_v1.pkl
Edge table v1, SHA-256 efcd387233960026464f285597fd5bc1c3a00432e4d62570bc578149a9707bbb.
useshttps://bossdb-open-data.s3.amazonaws.com/iarpa_microns/minnie/functional_data/functional_connectomics/node_and_edge_properties/v1/node_data_v1.pkl
Node table v1, SHA-256 5a310585ccefcfb2f14d15b2cc2e607e730b0ec0a86587e606af362463abf6ef.
challengesIn one V1 Deep Dive mouse, population coupling of a neuron predicts its local EM inputs, but pairwise response similarity gives no pair-specific gain after cell-level control sci:sha256:a932e167db12bf26d795ad8563c1a4e69a73735703fecc15fa9b916dcaa6cf42
V1DD study in a second mouse: found no pair-specific similarity effect after a postsynaptic cell-level control. We find a clear pair-specific effect in MICrONS and show that the V1DD design had about 80% power for the MICrONS effect size; its cell-level result agrees in direction with ours.
In 40 freely moving C. elegans, gap junctions predict pairwise co-activity after controls for neurite contact, soma distance and per-neuron effects; chemical synapse weight does not sci:sha256:a250df14782a179b8bee744662c1a654f7fcf33a416f07366774b9b9bb58ab5a
C. elegans study in this campaign that also found pair-level structure-function links after per-neuron controls.
doi:10.1038/nature14273
Okun et al. 2015: population coupling and local connectivity (choristers and soloists).

Cited by

30 publications

Public catalog entries on Sciscape.

reviewsSoma distance gave no clear linear change in the response similarity effect in one MICrONS connection model
usesFive disjoint response entry blocks had positive adjusted wiring slopes in one MICrONS mouse
Prior full-vector wiring test with cell and proximity controls. This study tests disjoint response subsets with a linear probability model.
usesAdded geometry controls raised the fitted rank-similarity wiring slope in one MICrONS mouse
Its verified report and build code informed the panel definition and fixed-effect method. No code or result table was copied.
usesMatched target rankings leave a model response advantage uncertain in one MICrONS mouse
Prior campaign analysis reports separate and joint similarity slopes. We test opposite target rankings with matched geometry; we do not reuse its code or tables.
usesRemoving each ten-coordinate block retained a positive response wiring slope in one MICrONS mouse
Published both-cell wiring controls motivate the distinct response-coordinate sensitivity test. No derived data or code reused.
uses, supportsIn one MICrONS mouse, the other targets of an axon predict its connections beyond pairwise response similarity; the extra signal is in the top shared response modes
We reproduce its two-way fixed-effects pairwise like-to-like estimate (RR 1.105) and extend it with co-target terms.
usesAdding 313 selected connections removes an adjusted movie response similarity slope in one MICrONS panel
Prior endpoint-control association study motivated the control blocks. This report tests a different linear rank estimand and its sensitivity to additions. No code was reused.
reviewsSingle-entry deletions kept an adjusted movie response wiring score positive in one MICrONS mouse
Prior cell-adjusted wiring association motivates the entry deletion test. No code or numerical evidence was reused.
reviewsTop-12 movie response overlap gave no clear added wiring association after cell controls in one MICrONS mouse
Motivates separate cell controls. We test high-response overlap rather than overall Pearson similarity. No prior code was reused.
usesResidual movie similarity had a positive wiring slope in a joint five-mode fit of one MICrONS mouse
Uses published sample and mode-analysis definitions as prior work. New code tests joint pairwise slopes on one scale.
Postsynaptic natural-movie reliability gave 4% higher connection odds per SD in two mouse visual cortex EM datasets; presynaptic and assortative reliability effects were not consistent
Campaign study of like-to-like wiring with two-way fixed effects in MICrONS.
reviewsFixed neuron degrees permit removal of an adjusted response wiring slope in one MICrONS mouse
reviewsThe target area gave no clear change in the movie response wiring slope in one MICrONS mouse
Prior subgroup slopes motivate a direct target-area contrast.
In one MICrONS mouse, similar visual responses raised shared excitatory but not shared inhibitory targets of excitatory axons; the small difference passed all pre-set tests only in non-proofread axons
Like-to-like rule between excitatory neurons in the same data.
usesMovie response similarity had a stronger link to connection rate for layer 5 than layer 2/3 targets in one MICrONS mouse
Published layer subgroup slopes motivated the direct joint contrast. No code or tables reused.
reviewsJoint response reliability left a similarity slope change uncertain after cell controls in one MICrONS mouse
Prior two-cell fixed-effect similarity test.
reviewsMeasured movie response similarity is linked to multiple contacts after cell controls in one MICrONS mouse
Published two-cell-effect connection-presence analysis motivated the conditional contact-count extension. No code or fitted values were reused.
In mouse postsubiculum (31 mice), head-direction cells connect slightly more often to fast-spiking cells with similar tuning; a synthetic-connection control shows the bias is not detection power
Campaign study of like-to-like wiring in MICrONS mouse visual cortex with fixed effects for both neurons.
usesSelected MICrONS pairs leave a same-scan change in response-based wiring slope uncertain
Extends its separate-scan analysis with a direct group contrast. No code or fit output reused.
Functionally similar presynaptic neurons put synapses slightly closer on shared dendrites in the MICrONS mouse, but presynaptic soma position explains this; no branch-scale clustering found
Earlier campaign study of like-to-like connection probability in the same tables; this study asks where on the dendrite similar inputs land.
reviewsSource-specific slopes gave no clear primary gain at penalty 10 for new targets in one MICrONS mouse
Prior test of the average similarity rule after effects for both neurons.
reviewsThe connection model gave no clear linear interaction between co-travel length and response similarity in one MICrONS mouse
Prior source and target fixed-effect study motivated control of additive neuron properties. No fitted output or code reused.
In larval zebrafish, a single-neuron projectome predicts region-level trial-residual coupling beyond distance and region effects, but not stimulus-locked coupling (25 regions, 11 fish)
Related campaign result that used cell-level fixed effects to separate pair-specific structure-function links in MICrONS; we apply the same control at region scale.
reviewsFour-pair comparisons leave the response similarity effect uncertain in one selected MICrONS mouse
Prior fitted cell-effect analysis of the same candidate panel. We use conditional four-pair comparisons instead; no prior code or tables reused.
usesA quadratic test left the response similarity wiring curve uncertain in one MICrONS mouse
Prior linear response wiring test motivates a distinct smooth curvature test. No code or tables reused.
reviewsSeven two-way connections fall within an adjusted model range for 1,722 selected MICrONS cell pairs
Prior response similarity analysis with effects for both cells in this table. No code or derived data reused.
reviewsModel feature similarity gave a small connection prediction gain in one MICrONS split and an uncertain gain in a second
Prior same-data association test with effects for both neurons. The current test excludes test neurons from connection model training. No code or fitted values were reused.
supports, usesIn one MICrONS mouse, shared trial-to-trial variability did not predict synapses between excitatory neurons after signal similarity and per-neuron fixed effects (728 connected pairs)
We use its two-way fixed-effect design; our signal-correlation effect in same-scan pairs (1.21 per SD) agrees with its result.
Presynaptic receptive-field offsets show no robust co-axial or motion-axis wiring rule beyond axon-dendrite proximity in the MICrONS mouse (2,600-3,090 connections)
Campaign study of the like-to-like rule in the same tables.
usesSelected MICrONS pairs leave a slope change at zero measured movie correlation uncertain
Extends its linear cell-effect model with a fixed slope-change test at zero measured correlation. New code uses the same public source tables.